=== alexisb-afk is now known as alexisb === MrBIOS_ is now known as MrBIOS === tikun is now known as sikun === House_ is now known as House [06:03] nacc: interesting. I guess we should fix the release then :-/ [06:19] ThiagoCMC: you can always twek things as needed via http://cloudinit.readthedocs.io/en/latest/topics/examples.html#run-commands-on-first-boot if not natively supported [06:19] ThiagoCMC: there is way more net config to come by the means of https://lists.ubuntu.com/archives/ubuntu-devel/2016-July/039464.html [06:35] ThiagoCMC: depending on what you are looking for you might look at http://cloudinit.readthedocs.io/en/latest/topics/datasources.html#no-cloud to inject e/n/i for now? [06:48] Hi, is it possible to copy a LXC1 container to a LXC2 host or is there a way to convert the LXC1 container? [07:08] innercode: you mean lxc1 to lxc2 and not to lxd right? [07:09] cpaelzer: That's it [07:10] innercode: that should be a straight upgrade on the host itself according to e.g. http://lxc-users.linuxcontainers.narkive.com/IsqEwMzt/upgrade-lxc1-to-lxc2-lxd [07:10] innercode: but you were thinking on moving off to a newer host right? [07:11] maybe making a clone of your current host, then do the upgrade on the clone and roll-back to the original in case things went wrong? [07:11] cpaelzer: Yes, but I can upgrade the current host first [07:11] if you can clone you can also use it for various verifications before doing the real switch [07:12] innercode: of course you "can" upgrade the current host first, just IMHO it is always wise to have a fallback strategy [07:12] cpaelzer: I will backup first the current host, upgrade lxc and then move it to the new host [07:12] cpaelzer: Thanks for your help and link [07:13] innercode: that sounds reasonable, only one change at a time and a way to fall back if needed [07:15] jamespage: it seems OVS gos into final apply&clenaup mode several important fixes I was already keeping a list to backport got applied the last two days [07:15] * cpaelzer is touching all kind of wood for a great ovs 2.6 release [07:32] cpaelzer, I'll look at another snapshot in advance of final release [07:42] ddellav, coreycb: doing designate rc1 [08:56] coreycb, hmm [08:56] coreycb, ddellav: designate is done [09:02] coreycb, ddellav: https://review.openstack.org/#/c/372146 [09:02] concerns me [09:02] ceilometer-api binary is actually broken in rc1 - it worked ok in b3 === iberezovskiy|off is now known as iberezovskiy [09:21] ddellav, coreycb: infact this is why ceilometer is stuck in proposed - its autopkgtests fail due to the broken ceilometer-api binary [09:43] jamespage: we got the first report on dpdk no more able to handle the permission/ownership fix [09:43] jamespage: I created bug 1625542 to track [09:43] bug 1625542 in dpdk (Ubuntu) "--vhost-owner --vhost-perm no more transferred to DPDK EAL commandline" [Undecided,Incomplete] https://launchpad.net/bugs/1625542 [09:44] jamespage: but I think we wait the few days until you picked up a OVS 2.6 and then add delta as (if) needed [09:44] there still is the small but existing chance that they pick up https://www.mail-archive.com/dev@openvswitch.org/msg69706.html [09:44] Good morning [09:46] Last week I was experiencing an issue with apt-get update taking an extremely long time via the AWS eu-west-1 package mirror, especially with a few instances doing it simultaneously [09:46] I'm doing some tests now and can't seem to replicate the problem I was having, so I wondered if there were any known issues with the package archive last week that have now been resolved? === sts is now known as Guest8435 === IdleOne is now known as Guest66320 === _degorenko|afk is now known as degorenko [12:26] jamespage, wth, dropped ceilometer-api? [12:26] in RC1 [12:26] coreycb, no its been replaced with an equivalent command that runs wsgi app standalone [12:26] I've picked the patch and updated the packaging [12:27] jamespage, ok so nothing major after all? I'll take a look at what you did. [12:27] coreycb, no it was a bit of a false alarm [12:27] the term wsgi script is overloaded - at least three people have asked the same question about the change [12:28] jamespage, ok [12:38] rbasak: interesting ... you might be able to explain me why I can assign bug tasks for Xenial on my own, but only nominate for P&T [12:48] jamespage, any idea what this means? E: Could not find python-mock-services*/python-mock-services_*.dsc http://10.245.168.2:8080/job/backport_package/1481/console [12:49] coreycb, hmm [12:50] coreycb, cloud-archive-backport: Cloud Archive is already up-to-date for python-mock-services. [12:50] that should stop it trying to build the package? [12:51] jamespage, oh duh. yeah I think it usually quits at that point. [12:51] bom does [12:51] not sure about individual backports [12:51] might need a tweak [12:51] jamespage, ok === Thorn__ is now known as Thorn [13:10] cpaelzer_: packagesets ACLs are release-specific. So it may be that there are things missing from the ACL for you for older releases. [13:56] coreycb, we need to discuss dnsmasq in UCA [13:56] coreycb, we managed not tod that last week [13:57] jamespage, yeah [13:57] jamespage, it seemed like a self-contained backport on my quick glance [13:58] coreycb, is this for ipv6 dnsmasq support [13:58] ? [13:59] jamespage, I think all we need is dhcp_release6 [13:59] jamespage, here's the neutron commit: https://review.openstack.org/#/c/301747/ [14:00] coreycb, ok +1 [14:00] lets make sure neutron gets a versioned dep as well please [14:02] jamespage, would you prefer to attempt an SRU of just the dhcp_release6 code to xenial? not sure if that would be acceptable since it's not a bug. [14:02] coreycb, hmm [14:02] can mitaka use it? [14:03] jamespage, looks like the neutron code is just in yakkety [14:03] jamespage, I mean, newton [14:03] coreycb, add it to the UCA then [14:03] its a feature so is unlikely to be SRU worthy [14:04] jamespage, ok [14:08] coreycb, adding swauth for newton UCA as well [14:09] jamespage, +1 [14:15] jamespage, dnsmasq backported successfully [14:15] jamespage, updating neutron d/control === Guest14517 is now known as med_ [14:33] is it possible to use ubuntu-vm-builder (or the just the underlying vm builder) via cli syntax to use a dedicated NIC? Like direct dev private macvtap? [14:33] or does that have to be done in a template at best? [14:46] Good morning all [14:46] (Or evening) [14:47] Does anyone have a ballpark estimate of how long the Raspberry Pi's Ubuntu version with a server pre-configured take to complete it's first boot? [14:47] Hi [14:47] Hey Voyage [14:47] Can anyone help with this Its a nightmare to bypass this: https://pastebin.mozilla.org/8911617 === JanC is now known as Guest12202 === JanC_ is now known as JanC [14:59] Voyage, you dont like being propmted for a passphase ? [14:59] what is the nightmare ? [15:04] smoser: how to I login ? [15:04] to what ? [15:05] itis asking you for the pass phrase to your ssh key. i dont know if your ssh key is present in the user that you're trying to get to or not. [15:05] aws [15:05] smoser: I changed computers [15:05] and used same ssh key [15:05] if you have public keys in the instance you launched, then [15:05] ssh ubuntu@that-ip-address [15:05] well, obviously the key is protected with a passphrase [15:06] and sshing in as 'root@that-ip-address' will tell you to do that. [15:07] RoyK: I never set it. I changed computers and used the same key. On this computer, I get a passphrase prompt that I cant bypass [15:09] debug1: Authentications that can continue: publickey <-- hm [15:10] not quite sure what you're doing here [15:11] lamont, bug 1611923. Do you need that back into xenial (and/or trusty) too? or just enough to fix in yakkety? [15:11] bug 1611923 in python-django (Ubuntu) "http.request does not support ipv6-formatted ipv4 addresses" [Medium,In progress] https://launchpad.net/bugs/1611923 [15:13] jgrimm: many things are relative... [15:13] jgrimm: we currently monkeypatch the hell out of it, so it doesn't absoltely need to be backported [15:13] OTOH, others might like it [15:14] lamont, certainly doable. i'll put it back to xenial at least then [15:14] jgrimm: so it's really a question of what our ipv6 support story is for xenial [15:14] trusty is dead to me (wrt ipv6) [15:14] too many breakages [15:15] lamont, fair enough. and why i was thinking xenial at least would be nice [15:15] yep [15:15] thanks sir [15:15] np [15:17] cpaelzer_, I was reading that very same doc! Thanks! The problem is that, if I disable the DHCP for a specific subnet, Cloud Init still configured the IP as static!!! But I do not want any setup for the secondarie networks. As a workaroung, I'm doing this on cloud init script: ifdown ens4 ; ifdown ens5 (while ens3 is fine, first). [15:18] ThiagoCMC: you might ask smoser in #cloud-init for an experts advise [15:21] ok, thanks! [15:21] is there a ubuntu kvm or libvirt sub group? [15:34] frickler, dnsmasq 2.76 and neutron that depens on it are working their way back to the newton cloud archive [15:34] depends [15:37] smoser: from a usability perspective, do you think it's a usability issue you hit yesterday (passing a non-xgit'd directory results in it starting from teh beginning). I can at least put that in the -h [15:39] nacc, i do think itunawell, i guess it just depends on who you think is going to run the importer [15:40] if the usd-importer branch is always up to date, its not a big deal. [15:40] if it is out of date, then i would really epxect that a normal user is probably wanting to git clone [15:40] and then 'usd-import' [15:40] smoser: yeah, i didn't optimize for regular users running it -- mostly myself in testing and then eventually the cronjob [15:40] smoser: yeah, but even that won't generally work even w/o xgit [15:40] as you have to have somewhere to pull sources too [15:41] but i'll have a think [15:42] nacc, i guess in my mind your time is better spent making the importer branch always up to date [15:42] then you wont have bothersome users saying "why doesnt it work like a bothersomeo user thinks it should" [15:44] nacc, a simple 'xgitify' would probably be sufficient too [15:44] dear user, if you want to run the importer your self do: [15:45] git clone http://.../your-package [15:45] usd-importer xgitify your-package your-xgit-dir [15:45] usd-import your-xgit-dir [15:45] where xgitify can basically just 'git clone' [15:45] smoser: yep, that makes sense [15:46] smoser: that's the thing, if you just use the importer (and don't clone), it does work :) [15:46] :) [15:47] thats a good point. [15:47] i was just trying to be too smart [15:47] and re-use what i had [15:50] yep [15:50] the other unfortunate bit i've hit is that git-clone and xgit are incompatible anyways [15:53] nacc, smoser: if enhancing xgit, it might be an idea to make it work with "git worktree" instead, now that it exists [15:54] Assuming that's possible [15:54] rbasak: yep, that's my theory [15:54] *thinking [15:54] we shouldn't "need" xgit beyond convenience [15:54] i also wonder if we could put the pull-*-source -d output in .git/junk or something [15:55] rbasak: ubuntu/devel should point to ubuntu/ if it exists and if not, then ubuntu/, correct? [15:56] As a type of cache? That could work well. [15:56] nacc: correct [15:57] rbasak: yeah, something like that, i'd need to dig into it -- i was thinking also as a means to not really need xgit itself (as it feels like the primary reason is to keep pulled things out of the working tree [15:57] nacc: agreed [15:58] rbasak: would then make smoser's git-clone case also 'just work', i think [16:00] rbasak, nacc get it importing everything on cron. and i will be uber happy. and never bother you again :) [16:00] smoser: :) we're getting very close to that [16:01] i think we've got a method to work around bad changelogs now, which seems to work well. I need to fix isc and then i think those are the two classes of major workarounds that we've had to recode for [16:22] rockstar, pylxd 2.0.5 is uploaded to xenial and I'll ask arges if he can review it. and nova-lxd b3 is uploaded to yakkety. [16:23] coreycb: ta [16:43] jgrimm: would you like to take bug 1595096? [16:43] bug 1595096 in postfix (Ubuntu) "cannot create multi postfix instance by postmulti command" [High,Triaged] https://launchpad.net/bugs/1595096 [16:44] looking [16:44] Looks like a cherry-pick from Debian for an SRU is all that is needed, though I haven't confirmed. [16:45] rbasak, looks like cpaelzer had done some looking at it === alexisb is now known as alexisb-afk [17:23] Hey! Anyone around tried using conjure-up to install openstack in single-machine mode around ? === iberezovskiy is now known as iberezovskiy|off [17:25] It's uh, not quite working out here, conjure-up itself is failing in a failcascade, and the more i get bleeding edge versions the worse it is, culminating at conjure-up just apt-get install 'ing argv0 in a loop. [17:26] argv0? o_O [17:26] well, i mean the conjure-up's arg [17:27] ah [17:27] that would be 1 not 0 [17:27] was all for the best though, I hadn't seen ssh-argv0 yet, heh [17:27] my bad [17:27] :P [17:28] so yeah the documentation looks pretty outdated, since the docs still speak of openstack-install which doesn't seem to exist anymore [17:29] tried going through conjure-up but it failed due to invalid LXD image names, so i tried taking the official image and naming it as requested, but it failed later on so i guess that wasn't quite it :P [17:30] tried getting conjure-up bleeding edge, and as i said, it just runs apt-get install in a loop with whatever you give it [17:30] which uh, is quite, err, weird [17:38] oh :) got it to work by beating it sufficiently with removal and reinstalls :p [17:38] the beatings will continue until morale improves [17:38] (sorry, I got nothing) === alexisb-afk is now known as alexisb === beisner is now known as beisner-food === beisner-food is now known as beisner [20:25] I'm trying to figure out why a crontab element isn't firing. I have one that is definitely working with the permissions -rwxrwxr-x and one that is not working with -rwxr-xr-x .. would that be the reason why? [20:34] is there a way to run a task from the cron user? to see if it works? [20:36] the user is not the only difference [20:44] jayjo: probably something makes an assumption about a PATH being set differently than it is. that's what it always is. :) [20:45] is there a way to echo some of those variables from the shell script? [20:45] /usr/bin/env > /tmp/environment_variables :) [20:58] PATH or another environment variable, trying to run some tool that expects a TTY, trying to access something with the wrong permissions, expecting a login session, etc. [21:48] docker: Error response from daemon: rpc error: code = 2 desc = "oci runtime error: flag provided but not defined: -bundle". [21:48] is ^ a known issue? [21:48] this is on a fresh Xenial install === MrBIOS_ is now known as MrBIOS